Introduction

You've probably seen centreoftheselights' AO3 Year In Review: 2025 (or one of the past years) floating around the internet before. Big color coded spreadsheet about the most tagged ships. It's never really been that interesting of a topic to me cause I generally don't give a crap about fandom trends (nor do I read literally any of the tags in the top100).

But despite not actually caring about the content, I find the choices made by centreoftheselights to be disagreeable when attempting to put together statistics like these. Especially in fandom where people care so much about these things (for whatever reason), providing incomplete statistics is something I disagree with quite a bit.

So this year I got annoyed at the lack of counting restricted works, the strange subtraction counting method, and decided to do it myself.

Despite that, I want to say that I have nothing against centreoftheselights as an individual, merely their methodology in data collection. This is not an attack, but a response. I wanted to look at this data myself, gather it in as complete a form as I could manage, and then compare the two to see how these methodologies differ, and by how much.

Yada yada don't harass people, be polite with criticism, etc etc. Enjoy the rest of the paper. Text version of the chart can be found at the bottom.

Reading the Chart

Generally I would say this is pretty self explanatory (and explained within the image as well, but here's each column explained:

Leftmost we have the ranking. This goes 1 to 139, and is ordered by total taggings.

Next is the comparison between a given rank and centreoftheselights' AO3 Year In Review: 2025 chart. A red number indicates that the tag has dropped that many spaces, while a green number indicates that the tag has risen that many spaces. A cyan NEW means that the tag in question was not present on the AO3 Year In Review.

Column 3 is the name of the Tag.

Column 4 is the fandom that the tag hails from. Thankfully I don't think any of these are crossover tags. Correct me if that's incorrect.

Column 5 is the type or category of the relationship. This is explained in a footnote at the bottom of the image, but I mostly judged this off of the most-used category for that tag if I was unfamiliar with the characters. A caveat here is that character genders are not set in stone, and I've certainly read Naruto/Fem!Sasuke or Sasuke/Fem!Naruto fics before. Despite that, these are a pretty tiny minority so I feel comfortable saying that generally a tag's category is accurate. "Other" is used for relationships where one or both of the participants is not of a set gender (or is nonbinary, as with the Forsaken - Roblox tag), Gen is for non-romantic relationships (indicated with an ampersand), and Multi is for relationships with more than two participants. Lastly, N/A is for the three non-relationship tags in the relationship tag category.

After the Type, we have the Taggings column. This is the total number of taggings, i.e. works that use this tag. This was collected while logged in, yesterday and today, and should be mostly accurate, barring any late deletions.

Tags w/o Locked is the same data but collected while logged out (in an incognito window). This results in a smaller number for ever single tag, as all of these have at least a few locked works.

Difference₁ is the difference between Taggings and Tags w/o Locked. Basically subtracting the latter from the former. This shows how many works are locked for each tag.

The last 3 columns in grey all refer to centreoftheselights' AO3 Year In Review: 2025 chart. First I have the number the centreoftheselights gives for every ship shared between our two charts. For example, 10,028 is how many fics they claim use the tag Evan "Buck" Buckley/Eddie Diaz (9-1-1 TV). The Difference₂ column takes that number and compares it to the overall number from the Taggings column to give the number less than the total that centreoftheselights got. Lastly, the % column takes that numerical difference and turns it into a percentage of the total.

Methodology

First and foremost, when data was collected, the following search filter was used: created_at: ["2025-01-01" TO "2025-12-31"]. This limits the results to stories published between the 1st of January 2025 and the 31st of December 2025. As a note, if a story was published and then deleted during this time, it will not be counted. Additionally, if a story was published backdated (e.g. for 2004) but that posting was done in, say, February 2025, it is counted towards the numbers for that ship. This is partially because the extra effort to separate every single backdated work didn't seem worth it, and also the numbers of backdated works are presumably quite small and don't affect the overall placements all that much.

To assemble the list of relationship tags I made use of three strategies. First and worst, I went to each rating, and looked at the list of relationship tags Ao3 automatically offers, which is the 10 most populated. By excluding tags from that list until the newest joiners had less than 1000 works, I was able to gather around 70 tags.

Next, I collected every tag from centreoftheselights' AO3 Year In Review: 2025 list as well as Randomist1031's 2025 100+ Ao3 Most Written Pairings (with both Public + Restricted Fiction). Together these gave me around 140 relationship tags. 

Lastly, because I wanted this list to be as complete as I could reasonably do given the constraints of Ao3's search operators and options, I went to the Tag Search, sorted by Relationship, Uses, Descending, and went through 15 pages of relationship tags, opening nearly all of them and checking to see how many uses they had. Does this guarantee that I have every >2000 work ship? No. Page 15 is mid-4000 taggings. However, the amount of rate limits I was getting was getting tiresome, so I gave up. Please feel free to let me know if there's any tags I didn't include, and I will update this fic.

Finally I was left with a list of 160ish ships, which, after dropping any ships with less than 2000 work taggings, gives the final list of the top 139 Ao3 Relationship Tags.

My understanding is that for at least two of these tags (James Potter/Lily Evans Potter  and Neuvillette/Wriothesley) there's been some form of fanwork spam in the past year. I have no knowledge on this, but felt it warranted mentioning. I have not adjusted the rankings in any way.

In regards to presentation, I've included the number of taggings in 2025, the number of taggings when not logged in (i.e. no restricted/locked works), and the number of taggings provided by centreoftheselights. I do this to demonstrate how far their data is from actual reality, as I feel that while people can explain why their data collection is flawed, that's not quite as effective as seeing that they're not counting >5000 Wolfstar fics published this year.

Most important takeaway: Everything was collected and collated by hand. Any errors are mine, so please leave a comment mentioning them and I'll try to get them fixed as soon as possible.

Differences from Other Lists

You may note a different #1 on this list versus every other variety of this list you've seen. I'm choosing to include every single Relationship Tag with over 2000 taggings (hopefully, at least. I don't think I'm missing any). This includes /reader tags and technically-relationships-but-weird tags like the Original Character tags, but it also includes Other Background Relationship(s) and More Relationship Tags to Be Added. I think that including these adds some extra neat information and doesn't harm the integrity of the list, since I'm not cutting it off at 100 anyways.

One thing to note is that the Original Character(s)/Original Character(s) tag is a metatag for all OC/OC relationships, and as such is encompassing of the numbers for all other OC/OC tags on this list.

In comparison to centreoftheselights' list, the following tags are not represented due to being below the 2000 taggings cutoff:

• Jiang Yan | Jiang Wulang/Shao Dong Jia
• Shauna Shipman/Jackie Taylor
• Josh Dun/Tyler Joseph
• Jax/Pomni (The Amazing Digital Circus)
• Mira/Zoey (KPop Demon Hunters)
• Tian Xuning/Zi Yu (b. 2002)
• Aotsuki Eito/Sumino Takumi
• Ma Boquan/Zhang Kangle
• Mira/Rumi (KPop Demon Hunters)
• Will Byers & Mike Wheeler

In comparison to Randomist1031's list, the following tags are not represented due to being below the 2000 taggings cutoff.

• Lestat de Lioncourt/Louis de Pointe du Lac
• Miya Atsumu/Sakusa Kiyoomi
• Armand/Daniel Molloy
• Rook/Emmrich Volkarin
• Neuvillette/Wriothesley (Genshin Impact)
• Jinx & Vi (League of Legends)
• Edwin Paine | Edwin Payne/Charles Rowland
• Tartaglia | Childe/Zhongli (Genshin Impact)
• Logan | Worst Wolverine (Deadpool Movies)/Wade Wilson
• Bill Cipher/Ford Pines

Brief Analysis

Since it seems to be tradition, here's brief stats from this thing.

Out of 139 tags, 83 are M/M, 16 are Gen, 16 are F/M, 12 are Other, 8 are F/F, and 3 are non-relationship relationship tags marked N/A.

The average number of taggings for a tag on this list is 4,310. This is helped by the cutoff being 2000 taggings lol. 

The average percentage that centreoftheselights' numbers are off by is 22.14%, with the largest being 80.56% (James Potter/Lily Evans Potter), or, disregarding that due to the spam, 70.81% with Castiel/Dean Winchester. Numerically, the largest difference is again Jilly with a difference of 12,101 taggings. Discounting that, it's Sirius Black/Remus Lupin, with a difference of 5,556 taggings.

If every single tag usage was for its own unique work, then this would account for 599,142 works published this year. That's about 23.18% of the total number of fics published in 2025! Of course, since lots of these tags appear alongside each other, we can estimate that it's a good bit less. Maybe around 15% of the total works, but that's just a guess.

Thoughts

Throughout this project I get so used to seeing numbers in the thousands as "small" because they don't make the cut, and then I go look at the total number of works published in a fandom I actually read this year, and it's like, 1499 (shoutout Worm). I'm pretty sure in the fandom I actually write for (Delve – SenescentSoul), we've published 6 fics all year. And like all year I've read maybe like, 157–170 fics total. Which is a lot of words, like several million words, but it's not that large of an actual number of stories.

So on one hand I'm very jealous because I wish I had as many fics to read as people who read in any of these fandoms, but on the other hand I already have enough to read without needing to look at 16,000 Jilly fics or whatever.

P.S. I want to ruin OC tags forever for you guys. Think about the standard of parenthetical pluralization on Ao3 for tags like Minor or Background Relationship(s) or Original Character(s)/Original Character(s). That's great. All of them follow that pattern, EXCEPT for Original Male Character/Original Male Character. When it's OFC/OMC the parenthetical pluralization is correct, it's exclusively the male/male version where it's inconsistent.

Thanks

Special shoutout to Reddit user SpaceTransmission for giving me the singular prompt to put this together even after it took me longer than other people.

Thanks also goes to Kenn for saving me from a possible embarrassing moment.

Additional thanks to my friends for giving this a brief lookover and suggesting a few wording changes.